I am having trouble with a set of copilot agent deployed in a live LMS environment for client usage. The agents have been embedded in an iframe on the LMS, and currently, when prompting the bots it gives an errors message "This agent is currently unavailable. It has reached its usage limit. Please try again later." When reviewing the agents in the co-pilot studio I am able to prompt it and it runs perfectly. The analytics show that we have not ran enough session to have reached usage limits.
My suspicion is that it could be an issue between Copilot studio and the OpenAI GPT4o model it's running, but cant be sure as OpenAI downtime detectors doesn't show any concrete evidence that its a backend API call issue.
Any idea as to why it would give the error message even when we are not running agents constantly or maxed out our usage limits - from what I can tell. Also, how does one go about fixing the issue?
to summarize : Copilot studio is bill by credit, not by session.
about "When reviewing the agents in the co-pilot studio I am able to prompt it and it runs perfectly" : , It's a developpement screen, so it's not related to the number of credit you still have.
The other option about limit is using too much the agent, like you said there are "hidden" throttling limit depending on you licence model, here are the limit
Check if you are not mismatching what is a session and what is credit : credit consumption could go very fast (a tenant grounded query is abuot 12 credits)
Hope this help you :) if it answer please declare the answer as the good one, it's important for the community :)
Thank you for your reply! To ensure a great experience for everyone, your content is awaiting approval by our Community Managers. Please check back later.